OpenAI announced GPT-6 Astra, a new frontier model aimed at advanced agentic work and computer-use tasks. The company made the announcement after an outage earlier in the day, as first reported by Engadget.
OpenAI called GPT-6 Astra “the most intelligent and aligned model in the world.” It described the system as especially capable at work handled by software agents, which take actions on a user’s behalf, along with tasks that require operating a computer interface.
GPT-6 Astra targets software agents that take action
Agentic AI has become a central focus for model developers because it pushes systems beyond answering prompts. An agent may need to interpret a goal, work through several steps and interact with software while it completes a task. Computer-use capability refers to a model’s ability to work with a computer environment rather than simply generate text or code.
OpenAI’s latest model arrives as AI companies compete to make assistants useful for longer, more practical workflows. The emphasis matters because tools that can act for users raise the stakes for reliability and alignment: mistakes can affect files, software settings and work completed inside other services, rather than remaining confined to a chat response.
The company’s use of the term frontier model signals that it sees Astra as part of its highest-capability model line. That claim will draw attention from developers and businesses evaluating AI agents for tasks that currently require staff to move between applications, follow instructions and verify each step.
Users cannot access the model yet
GPT-6 Astra was not available to users when it was announced. That leaves OpenAI’s claims about its performance untested by the wider developer community, which often identifies practical limits in computer-control systems once they reach real-world products.
OpenAI has not provided a release date, identified the products or subscription plans that will offer access, or published independent benchmark results for Astra. The company also did not explain the cause or scope of the outage that preceded the announcement. Those details will determine when developers can assess the model and whether its agent capabilities reach consumer tools, enterprise products or both.
This article was produced with AI assistance from multi-source reporting and is published under our editorial standards.
